RTL languages subtitles show punctuation marks in reverse

[dorsharon]

On the AndroidTV version of Plex, RTL subtitles (Hebrew, Arabic and more..) show punctuation marks at the start of sentences instead of at the end.

It'd be great if there was a setting to auto-reverse all downloaded RTL subtitles.

[pannal]

Wouldn't that normally be something the subtitle creator should be doing?

I'm not at all familiar with RTL languages, so cut me some slack here please :)

Can you give me an example? Is the error inside the subtitle itself or does the Plex client behave wrong?

[dorsharon]

No, the subtitle creator is doing what he should do, creating a sentence like this:

שלום לך.

However, the Plex decodes it as:

.שלום לך

which puts the period on the start instead of the end.
Other players will play Heb subs just fine. It's only Plex that doesn't.

[dorsharon]

If you'd like an example, you can download SubtitleEdit 3.5, where there's a feature under Tools for Reverse RTL:

[pannal]

So the order of the characters in שלום לך. is already correct, and in .שלום לך it is as well, just the full stop should be at the end?

If you read from right to left, why wouldn't the full stop be on the left?

[dorsharon]

The characters are in the right order, it's only punctuation marks that are supposed to appear at the end of sentences that are not right.

Most players usually recognize RTL languages and simply render the subtitles so that the period, commas etc. appear correctly.
Plex considers the heb subs as LTR and so doesn't reverses the punctuation.

[dorsharon]

So the order of the characters in שלום לך. is already correct, and in .שלום לך it is as well, just the full stop should be at the end?

That's correct

[pannal]

Let me get this straight: ש is the first word/character you'd read in this example?

[pannal]

Ah, you write/read from right to left, but the punctuation still occurs at the right, not the left?

[dorsharon]

ש us the first character I'd read.
We read right to left, Punctuation occurs on the left.

[pannal]

Wait, when punctuation occurs on the left, then .שלום לך is correct if I understood you fine.

[dorsharon]

Yeah. Github actually also reversed my punctuation for my examples.
Bottom line is that period should occur on the left, Plex shows it on the right on AndroidTV

[pannal]

OK. Is that the case with SZ subtitles only? I'm doing quite a bit of cleanup so I might kill the RTL marker in the process. Can you try downloading a subtitle manually from OpenSubtitles and see whether it really is Android TV which does that?

[dorsharon]

I downloaded from OpenSubtitles and the subtitles there are fine.
Furthermore, on other platforms Plex plays the subtitles fine, It's the Android TV version that doesn't.

This issue isn't exclusive to SZ, I never said that it was.
I was simply asking for a feature of a setting to auto-reverse these punctuation marks, so that Israeli users can finally fully enjoy SZ, as well as arabic and others.

[pannal]

This issue isn't exclusive to SZ, I never said that it was.

And I didn't say you did. I wanted to have as much information about the issue as possible, before I try and fix things :)

Edit: Can you attach an example subtitle so I can play around with it?

[dorsharon]

Vikings.S05E03.720p.HDTV.x264-KILLERS.B291968.msg1.zip

[pannal]

Please try the attached srt file. I've added an explicit RTL mark right after the first dot on:

4
00:00:08,230 --> 00:00:09,940
נראה שאתה לא מתייעץ איתנו
.‏בשום דבר

Does that work? If so, this modification could be very easy.

testout.zip

Edit: If not, the modifcation may be a bit more complex.

[dorsharon]

I'll check later today

[pannal]

Any news?

[dorsharon]

Sorry about the delay, been crazy busy.
Unfortunately this still doesn't change anything. The problem persists.

[pannal]

Damn. I'm not sure on how to approach this to be honest.
Your client seems to put the dot to the right side of the sentence explicitly, while the dot already being in the correct place (left side) inside the subtitle. I currently have no idea on how to circumvent that - have you actually had success in manually modifying a subtitle so it gets played correctly on your TV client?

Edit: I should have a look at what SubtitleEdit does.

[dorsharon]

Yes, SubtitleEdit's feature does the job.
Just annoying having to do it every single time...

[pannal]

OK I'm still not 100% sure why the SubtitleEdit approach works, simply because I have no experience with RTL languages. I think what it does it actually just putting the punctuation to the end of the sentence, explicitly, while keeping the rest intact.

Why that fixes things for you I still don't know, but if that's what you need, I can integrate it into SZ.

[dorsharon]

Great! Looking forward to testing it out :)

[pannal]

I think it'd be good if I separate the logic of the common modification for RTL and LTR languages, as they're currently only tested on LTR and probably won't work correctly on LTR. What do you think?

[dorsharon]

Sounds about right

[pannal]

Any update?

[dorsharon]

Hi, sorry that I've been away. Life got in the way big time.
Anyway, I want to revisit this issue for a possible fix.
What do you need from me to continue?

[pannal]

Can you re-download https://github.com/pannal/Sub-Zero.bundle/tree/heb_test? It should swap the dashes now, as well.

[pannal]

News?

[dorsharon]

How do I mod the subs with your heb-fix?
I've installed your version.

[pannal]

The same as before, go to an item in question using the SZ plugin menu, modify subtitle, Reverse punctuation in RTL languages.

[dorsharon]

Yeah that solves the dashes as well :)

[pannal]

Oh so that fix is working?

Do you want this to be a configurable default mod, that always gets applied to new subtitles, if enabled, or should I keep it manual?

[dorsharon]

Enabled by default, can be disabled manually

[pannal]

Well you will be able to enable it in SZ's settings, so it's applied by default. Then you can remove it for a specific item by going to the same menu you just went, and remove the mod.

[dorsharon]

Fantastic

[pannal]

I'll only enable this mod for Hebrew right now. If there are other languages where this applies, please give me a list.

[dorsharon]

Alright

[pannal]

This is properly implemented in this branch right now: https://github.com/pannal/Sub-Zero.bundle/tree/develop-2.5

And will be in the upcoming release. You can switch to that branch, it has the option to enable the modification by default (via SZ's settings).

[doopler]

Hi this is a great add on to already great plugin. but i have noticed that not all punctuation marks are reversed. Can you add comma (,) and other marks like (:), ("), (-) this?
thank you

[Zuerrex]

Hello there,

Please note that Arabic suffer from the same issue as well. It will be great if you could fix Arabic subtitles.

The below languages are RTL but I cannot confirm if those languages suffer from the same issue:

• Aramaic
• Azeri
• Dhivehi/Maldivian
• Kurdish (Sorani)
• Persian/Farsi
• Urdu

Thanks

[AnewAlgerianboy]

Please try the attached srt file. I've added an explicit RTL mark right after the first dot on:

4
00:00:08,230 --> 00:00:09,940
נראה שאתה לא מתייעץ איתנו
.‏בשום דבר

Does that work? If so, this modification could be very easy.

testout.zip

Edit: If not, the modifcation may be a bit more complex.

Hi Pannal ,And Thank You For Your Help.
For Arabic It Work's ,Take A Look Your-Self

Before

After

But Am Really NOOB At This Stuff
Can You Please Teach Me How To Do It

And What I Did In The Subtitle In The Photos Is I Paste The Subtitle Text In The File You Put In You Reply

[adirt4289]

any updates on this?

[pannal]

No, this plugin is discontinued.

[adirt4289]

No, this plugin is discontinued.

so no sulotion for this problem?

[pannal]

No, maybe try bazarr.